MVP 2

Working algorithm

Yes

it works

somehow

built for Linux and macOS

should work on Windows as well

(as soon as we figure out portaudio and libsndfile)

Original schedule

https://github.com/straend/MMAI-reverb/wiki/Schedule

Delay line?

double process_delay(delay_line_s *dl, double x)
{
    double y = dl->delay[dl->ptr];
    dl->delay[dl->ptr++] = x + (y * dl->gain);

    if (dl->ptr >= dl->delay_samples) {
        dl->ptr -= dl->delay_samples;
    }
    return y;
}

Tap delays, Boston symphony hall

const double ER_TAPS[18] = {
    0.0043, 0.0215, 0.0225, 0.0268, 0.0270, 0.0298, 0.0458, 0.0485, 0.0572, 
    0.0587, 0.0595, 0.0612, 0.0707, 0.0708, 0.0726, 0.0741, 0.0753, 0.0797
};
const double ER_GAINS[18] = {
    0.841, 0.504, 0.491, 0.379, 0.380, 0.346, 0.289, 0.272, 0.192, 
    0.193, 0.217, 0.181, 0.180, 0.181, 0.176, 0.142, 0.167, 0.134
};

Thank's, Moorer

https://github.com/keithhearne/VSTPlugins/blob/master/MoorerReverb/Source/ERTapDelayLine.h

and Allpass?

double process_allpass(delay_line_s *dl, double x)
{
    double y = dl->delay[dl->ptr];
    dl->delay[dl->ptr++] = 
        (dl->gain + (x + (y * dl->gain))) / 
        (1 + (dl->gain * (x + (y * dl->gain))));

    if (dl->ptr >= dl->delay_samples) {
        dl->ptr -= dl->delay_samples;
    }
    return y;
}

how about a Comb?

double process_comb(delay_line_s *dl, double x)
{
    double y = dl->delay[dl->ptr];
    dl->delay[dl->ptr++] = 
        (dl->gain) / 
        (1 + (dl->gain * (x + (y * dl->gain))));

    if (dl->ptr >= dl->delay_samples) {
        dl->ptr -= dl->delay_samples;
    }
    return y;
}

Everything

void try_moorer(double *samples, SF_INFO *sfinfo)
{
  double *early_reflections
      =calloc(sizeof(double), sfinfo->channels*sfinfo->frames);
  memccpy(early_reflections, samples, 
      sfinfo->channels*sfinfo->frames, sizeof(double)
  );
  just_delays(early_reflections, sfinfo);

  double *late_reflections
      =calloc(sizeof(double), sfinfo->channels*sfinfo->frames);
  memccpy(late_reflections, early_reflections, 
      sfinfo->channels*sfinfo->frames, sizeof(double)
  );

  comb_filters(late_reflections, sfinfo);
  allpass(late_reflections, sfinfo);
  double dry=0.3;
  double wet=1-dry;

  for(uint32_t i=0; i<sfinfo->channels*sfinfo->frames;i++){
    samples[i] = samples[i] * dry + late_reflections[i]*wet;
  }
  free(early_reflections);
  free(late_reflections);
}

lots of text

no audio

Audio("../audio/Epic Sax Guy.wav")

Audio("../build/epic.wav")

That worked great...

Lets use the demo effect instead

Live demo is best demo

note to selfie: run demo in Terminal